Accurate modeling of non-covalent interactions involving sulfur today is ubiquitous, particularly with regards to the role played by sulfur-containing heterocycles in the field of organic electronics. The density functional tight binding (DFTB) method offers a good compromise between computational efficiency and accuracy, enabling the treatment of thousands of atoms at a fraction of the cost of density functional theory (DFT) evaluations. DFTB is an approximate quantum chemical approach that is based on the DFT total energy expression. Here we address a critical issue inherent to the DFTB parameterization, which prevents the use of the DFTB framework for simulating non-covalent interactions involving sulfur atoms and precludes its combination with a dispersion correction.1–5 Dramatic examples of structural patterns relevant to the field of organic electronics illustrate that DFTB delivers erroneous (i.e., qualitatively wrong) results involving spurious binding.